pairkat

PaIRKAT assesses statistical relationships between networks of metabolites (pathways) and an outcome of interest (phenotype) by incorporating graph topography into kernel machine regression.

Key Features:

• KEGG Database Integration: PaIRKAT queries the KEGG database to determine interactions between metabolites from which network connectivity is constructed.
• Graph Topography Incorporation: The model framework improves testing power on high dimensional data by including graph topography in the kernel machine regression setting.
• Semi-Parametric Kernel Machine Regression: Provides a framework for testing relationships between outcomes of interest and high dimensional data such as metabolomic, genomic, or proteomic pathways.
• Network Representation: Uses known biological connections between high dimensional variables by representing them as edges of graphs or networks.
• Graph Regularization: Includes a graph regularization or smoothing approach for managing disconnected nodes within the graph.

Scientific Applications:

• Metabolomic Pathway Analysis: Testing for relationships between phenotypes and networks of metabolites.
• Genomic Pathway Analysis: Testing for relationships between outcomes and high dimensional genomic pathways.
• Proteomic Pathway Analysis: Testing for relationships between outcomes and high dimensional proteomic pathways.

Methodology:

PaIRKAT uses semi-parametric kernel machine regression with graph topography incorporated into the kernel function, queries the KEGG database to construct metabolite interaction networks, and applies graph regularization or smoothing to handle disconnected nodes.